Budget or throttle mechanism for node expirations

[Beardface123]

As a platform administrator we require that nodes do not exceed a certain age, which is an ideal use case for the NodePools expireAfter setting. Unfortunately, if there is a large influx of nodes to address pod scheduling demand, there is a fair chance all of those nodes will be terminated at the same time.

Disruption budgets are a very intuitive tool Karpenter implemented to controlled disruption for most cases in a cluster, however they do not handle this case where too many nodes reach their expireAfter period, all at once. The documented way to handle this would be PodDisruptionBudgets but this becomes a process and people issue in large multi-tenant clusters.

As an platform administrator, we will advocate for PodDisruptionBudgets to be used appropriately, however we'd also like to see a means to regulate or control how many nodes are terminated concurrently due to them expiring, much like we can when they are Drifted, Underutilized or Empty.